M4 Pro (20-core GPU, 24 GB)
3 records across 3 models (quants: Q4_K - Medium).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M4 Pro (20-core GPU, 24 GB) | llama-3-1-8b-instruct | Q4_K - Medium | llamafile | 32.533236711111115 | 359.5577164 | – | 3.419087840755556 | – | – | sourced | – | source |
| M4 Pro (20-core GPU, 24 GB) | llama-3-2-1b-instruct | Q4_K - Medium | llamafile | 119.24874207407407 | 2128.7541195555555 | – | 0.5725979937407406 | – | – | sourced | – | source |
| M4 Pro (20-core GPU, 24 GB) | qwen-2-5-14b-instruct | Q4_K - Medium | llamafile | 17.982970555555557 | 190.44284311111113 | – | 6.6294910853968245 | – | – | sourced | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085